MDEV-6776 ujis and eucjmps erroneously accept 0x8EA0 as a valid byte …

…sequence
MariaDB · Sep 24, 2014 · 3416fac · 3416fac
1 parent 9fa62b4
commit 3416fac
Show file tree

Hide file tree

Showing 6 changed files with 90 additions and 18 deletions.
diff --git a/mysql-test/r/ctype_eucjpms.result b/mysql-test/r/ctype_eucjpms.result
@@ -33622,3 +33622,20 @@ hex(weight_string(cast(0x8FA2C38FA2C38FA2C3 as char),25, 4, 0xC0))
 #
 # End of 5.6 tests
 #
+#
+# Start of 10.0 tests
+#
+#
+# MDEV-6776 ujis and eucjmps erroneously accept 0x8EA0 as a valid byte sequence
+#
+CREATE TABLE t1 (a VARCHAR(10) CHARACTER SET eucjpms);
+INSERT INTO t1 VALUES (0x8EA0);
+SELECT HEX(a), CHAR_LENGTH(a) FROM t1;
+HEX(a)	CHAR_LENGTH(a)
+	0
+DROP TABLE t1;
+SELECT _eucjpms 0x8EA0;
+ERROR HY000: Invalid eucjpms character string: '8EA0'
+#
+# End of 10.0 tests
+#
diff --git a/mysql-test/r/ctype_ujis.result b/mysql-test/r/ctype_ujis.result
@@ -25928,3 +25928,20 @@ hex(weight_string(cast(0x8FA2C38FA2C38FA2C3 as char),25, 4, 0xC0))
 #
 # End of 5.6 tests
 #
+#
+# Start of 10.0 tests
+#
+#
+# MDEV-6776 ujis and eucjmps erroneously accept 0x8EA0 as a valid byte sequence
+#
+CREATE TABLE t1 (a VARCHAR(10) CHARACTER SET ujis);
+INSERT INTO t1 VALUES (0x8EA0);
+SELECT HEX(a), CHAR_LENGTH(a) FROM t1;
+HEX(a)	CHAR_LENGTH(a)
+	0
+DROP TABLE t1;
+SELECT _ujis 0x8EA0;
+ERROR HY000: Invalid ujis character string: '8EA0'
+#
+# End of 10.0 tests
+#
diff --git a/mysql-test/t/ctype_eucjpms.test b/mysql-test/t/ctype_eucjpms.test
@@ -520,3 +520,23 @@ set collation_connection=eucjpms_bin;
 --echo #
 --echo # End of 5.6 tests
 --echo #
+
+
+--echo #
+--echo # Start of 10.0 tests
+--echo #
+
+--echo #
+--echo # MDEV-6776 ujis and eucjmps erroneously accept 0x8EA0 as a valid byte sequence
+--echo #
+CREATE TABLE t1 (a VARCHAR(10) CHARACTER SET eucjpms);
+INSERT INTO t1 VALUES (0x8EA0);
+SELECT HEX(a), CHAR_LENGTH(a) FROM t1;
+DROP TABLE t1;
+--error ER_INVALID_CHARACTER_STRING
+SELECT _eucjpms 0x8EA0;
+
+
+--echo #
+--echo # End of 10.0 tests
+--echo #
diff --git a/mysql-test/t/ctype_ujis.test b/mysql-test/t/ctype_ujis.test
@@ -1349,3 +1349,23 @@ set collation_connection=ujis_bin;
 --echo #
 --echo # End of 5.6 tests
 --echo #
+
+
+--echo #
+--echo # Start of 10.0 tests
+--echo #
+
+--echo #
+--echo # MDEV-6776 ujis and eucjmps erroneously accept 0x8EA0 as a valid byte sequence
+--echo #
+CREATE TABLE t1 (a VARCHAR(10) CHARACTER SET ujis);
+INSERT INTO t1 VALUES (0x8EA0);
+SELECT HEX(a), CHAR_LENGTH(a) FROM t1;
+DROP TABLE t1;
+--error ER_INVALID_CHARACTER_STRING
+SELECT _ujis 0x8EA0;
+
+
+--echo #
+--echo # End of 10.0 tests
+--echo #
diff --git a/strings/ctype-eucjpms.c b/strings/ctype-eucjpms.c
@@ -67418,10 +67418,10 @@ my_wc_mb_eucjpms(CHARSET_INFO *cs __attribute__((unused)),
 
 /*
   EUCJPMS encoding subcomponents:
-  [x00-x7F]			# ASCII/JIS-Roman (one-byte/character)
-  [x8E][xA0-xDF]		# half-width katakana (two bytes/char)
-  [x8F][xA1-xFE][xA1-xFE]	# JIS X 0212-1990 (three bytes/char)
-  [xA1-xFE][xA1-xFE]		# JIS X 0208:1997 (two bytes/char)
+  [x00-x7F]                     # ASCII/JIS-Roman (one-byte/character)
+  [x8E][xA1-xDF]                # half-width katakana (two bytes/char)
+  [x8F][xA1-xFE][xA1-xFE]       # JIS X 0212-1990 (three bytes/char)
+  [xA1-xFE][xA1-xFE]            # JIS X 0208:1997 (two bytes/char)
 */
 
 static
@@ -67444,15 +67444,15 @@ size_t my_well_formed_len_eucjpms(CHARSET_INFO *cs __attribute__((unused)),
     if (b >= (uchar *) end)         /* need more bytes */
       return (uint) (chbeg - beg);  /* unexpected EOL  */
 
-    if (ch == 0x8E)                 /* [x8E][xA0-xDF] */
+    if (iseucjpms_ss2(ch))          /* [x8E][xA1-xDF] */
     {
-      if (*b >= 0xA0 && *b <= 0xDF)
+      if (iskata(*b))
         continue;
       *error=1;
       return (uint) (chbeg - beg);  /* invalid sequence */
     }
 
-    if (ch == 0x8F)                 /* [x8F][xA1-xFE][xA1-xFE] */
+    if (iseucjpms_ss3(ch))          /* [x8F][xA1-xFE][xA1-xFE] */
     {
       ch= *b++;
       if (b >= (uchar*) end)
@@ -67462,8 +67462,7 @@ size_t my_well_formed_len_eucjpms(CHARSET_INFO *cs __attribute__((unused)),
       }
     }
 
-    if (ch >= 0xA1 && ch <= 0xFE &&
-        *b >= 0xA1 && *b <= 0xFE)   /* [xA1-xFE][xA1-xFE] */
+    if (iseucjpms(ch) && iseucjpms(*b)) /* [xA1-xFE][xA1-xFE] */
       continue;
     *error=1;
     return (size_t) (chbeg - beg);    /* invalid sequence */

diff --git a/strings/ctype-ujis.c b/strings/ctype-ujis.c
@@ -203,10 +203,10 @@ static uint mbcharlen_ujis(CHARSET_INFO *cs __attribute__((unused)),uint c)
 
 /*
   EUC-JP encoding subcomponents:
-  [x00-x7F]			# ASCII/JIS-Roman (one-byte/character)  
-  [x8E][xA0-xDF]		# half-width katakana (two bytes/char)  
-  [x8F][xA1-xFE][xA1-xFE]	# JIS X 0212-1990 (three bytes/char)  
-  [xA1-xFE][xA1-xFE]		# JIS X 0208:1997 (two bytes/char)
+  [x00-x7F]                        # ASCII/JIS-Roman (one-byte/character)  
+  [x8E][xA1-xDF]                   # half-width katakana (two bytes/char)  
+  [x8F][xA1-xFE][xA1-xFE]          # JIS X 0212-1990 (three bytes/char)  
+  [xA1-xFE][xA1-xFE]               # JIS X 0208:1997 (two bytes/char)
 */
 
 static
@@ -231,15 +231,15 @@ size_t my_well_formed_len_ujis(CHARSET_INFO *cs __attribute__((unused)),
       return (size_t) (chbeg - beg);            /* unexpected EOL  */ 
     }
 
-    if (ch == 0x8E)                 /* [x8E][xA0-xDF] */
+    if (isujis_ss2(ch))            /* [x8E][xA1-xDF] */
     {
-      if (*b >= 0xA0 && *b <= 0xDF)
+      if (iskata(*b))
         continue;
       *error= 1;
       return (size_t) (chbeg - beg);  /* invalid sequence */
     }
 
-    if (ch == 0x8F)                 /* [x8F][xA1-xFE][xA1-xFE] */
+    if (isujis_ss3(ch))           /* [x8F][xA1-xFE][xA1-xFE] */
     {
       ch= *b++;
       if (b >= (uchar*) end)
@@ -249,8 +249,7 @@ size_t my_well_formed_len_ujis(CHARSET_INFO *cs __attribute__((unused)),
       }
     }
 
-    if (ch >= 0xA1 && ch <= 0xFE &&
-        *b >= 0xA1 && *b <= 0xFE)   /* [xA1-xFE][xA1-xFE] */
+    if (isujis(ch) && isujis(*b)) /* [xA1-xFE][xA1-xFE] */
       continue;
     *error= 1;
     return (size_t) (chbeg - beg);    /* invalid sequence */